• No results found

Oil price shocks: Demand vs Supply in a two-country model

N/A
N/A
Protected

Academic year: 2021

Share "Oil price shocks: Demand vs Supply in a two-country model"

Copied!
41
0
0

Loading.... (view fulltext now)

Full text

(1)

MNB WORKING

PAPERS

2008/5

ALESSIA CAMPOLMI

Oil price shocks: Demand vs Supply

in a two-country model

(2)
(3)

Oil price shocks: Demand vs Supply

in a two-country model

(4)

Published by the Magyar Nemzeti Bank Szabadság tér 8–9, H–1850 Budapest

http://www.mnb.hu

ISSN 1585 5600 (online)

The MNB Working Paper series includes studies that are aimed to be of interest to the academic community, as well as researchers in central banks and elsewhere. Starting from 9/2005, articles undergo a refereeing process, and their

publication is supervised by an editorial board.

The purpose of publishing the Working Paper series is to stimulate comments and suggestions to the work prepared within the Magyar Nemzeti Bank. Citations should refer to a Magyar Nemzeti Bank Working Paper. The views

expressed are those of the authors and do not necessarily reflect the official view of the Bank.

MNB Working Papers 2008/5

Oil price shocks: Demand vs Supply in a two-country model* (Olajársokkok: a kereslet, illetve a kínálat hatása egy kétországos modellben)

Written by: Alessia Campolmi**

* I would like to thank Jordi Galí for excellent supervision. I am also thankful to Michael Reiter, Thijs van Rens, and all seminar participants at Universitat Pompeu Fabra, Magyar Nemzeti Bank and Central European University and conference participants at the “4th Annual Vienna Macroeconomic Workshop on Current Topics in Macroeconomic Theory and Policy” and ASSET Meeting 2007. A special thank to Stefano Gnocchi and Chiara Forlati for many useful discussions.

** CENTRAL EUROPEAN UNIVERSITY AND MAGYAR NEMZETI BANK. EMAIL: CAMPOLMIA@CEU.HU. PERSONAL HOME PAGE: HTTP://WWW.PERSONAL.CEU.HU/DEPARTS/PERSONAL/ALESSIA_CAMPOLMI/

(5)

Abstract

6

1 Introduction

7

2 Related literature

10

3 Stylized facts

12

4 The model

14 Household problem 14

Firm problem - Final goods sector 16

Firm problem - Intermediate goods sector 16

Equilibrium conditions 17

Flexible prices and wages 19

Sticky prices and wages 20

5 Demand vs Supply Shock: Impulse Response Analysis

22

Baseline Calibration 22

Demand vs Supply Shock 22

6 Conclusions

25

Appendices

26

Appendix A 26

Appendix B 28

(6)

Abstract

From the last quarter of 2001 to the third quarter of 2005 the real price of oil increased by 103%. Such an increase is comparable to the one experienced during the oil shock of 1973. At the same time, the behaviour of real GDP growth, Consumer Price inflation (CPI inflation), GDP Deflator inflation, real wages and wage inflation in the U.S. in the 1970s was very different from the one exhibited in the 2000s. What can explain such a difference? Within a two-country framework where oil is used in production, two kinds of shocks are analyzed: (a) a reduction in oil supply, (b) a persistent increase in foreign productivity (as proxy for the experience of China in the last years). It is shown that, while the 1970s are consistent with a supply shock, the shock to foreign productivity generates dynamics close to the one observed in the 2000s.

JEL Classification: E12, F41.

Keywords:oil price, open economy, demand and supply shocks.

Összefoglalás

2001 utolsó és 2005 harmadik negyedéve között az olaj ára reálértékben 103%-kal n˝ott. Ez nagyjából az 1973-as olajársokkal megegyez˝o emelkedés. Ugyanakkor az USA gazdasági mutatói (a reál GDP és a reálbérek növekedése, CPI infláció, a GDP deflátor alakulása, a bérinfláció) igen eltér˝oen alakultak a két id˝oszakban. Mi magyarázhatja ezt a különbséget? A tanulmány ehhez egy olyan kétországos modellt tekint, ahol az olaj is szerepel mint termelési tényez˝o. Két sokk hatását hasonlítjuk össze: (a) az olajkínálat csökkenése, (b) a külföldi ország permanens termelékenység-növekedése (a kínai gazdaság elmúlt id˝oszakbeli teljesítményét megjelenítend˝o). Megmutatjuk, hogy míg a 70-es években látott viselkedés az el˝obbi típusú, kínálati sokk hatásával írható le, addig a 2000-es évek történéseit jól magyarázza az utóbbi típusú, küls˝o termelékenységi sokk.

(7)

1

Introduction

"China’s oil demand certainly shot up in 2004, by 15%. That strong growth, along with political crises in Iran and Nigeria, appeared to explain why oil prices have been recently heading towards $70 a barrel again1".

The oil price went from 27$ per barrel in January 2000 to 74$ per barrel in July 2006 and now is well above the 100$ (125$ in May 2008)2. Such an increase in the price of oil is of an even bigger magnitude than the one experienced during the 70s. But, is thisnew oil shockcomparable with the one in the 70s? While in the 70s the rise in oil price was accompanied by a reduction in real GDP and an increase in inflation, the situation in the 2000s is quite different with no signs of either strong inflationary pressures or output slowdown3. When studying oil price changes, no attention is commonly dedicated to the source of the oil price increase. The implied assumption under this approach is that the only thing that matters for the economy is the sign (and the magnitude) of the oil price change. The contribution of the paper is twofold: first, to show the failure of this approach in explaining the differences between the oil price changes of the 70s and the current one; second, to propose and alternative approach where a "step back" with respect to the current literature is taken. Specifically, we develop a model where shocks of different nature can drive an increase in the oil price and we show that while the 70s are consistent with a supply shock, a demand shock delivers dynamics close to the one observed in the 2000. The policy implication of this result is a clear one: to uncover which shocks are the sources of the current rise in oil price is fundamental to understand whether or not the economy is at risk of experiencing a period of recession and high inflation like in the 70s.

The intuition of why different sources of oil price increase can deliver different dynamics is a simple one. Consider an exogenous reduction in the supply of oil. This experiment is consistent with the hypothesis commonly used in the literature to study the oil shocks of the 70s4. As a consequence, the economy will experience an increase in oil price which will boost marginal costs therefore driving an increase in inflation (both in prices and wages), and a decrease in real wages and GDP. While this is consistent with the experience of the 70s, it is clearly at odds with what we observe today. Consider instead a two country model where an exogenous and persistent increase in the productivity of the foreign country drives an increase in foreign GDP. If this shock is very persistent (like it seems to be the case for a country like China), then several things will happen at the same time. First, given the increase in production, the foreign country will increase the oil demand therefore driving up the oil price. On the home economy this will have the same consequences of an oil price increase driven by a reduction in oil supply, i.e. marginal costs will increase therefore domestic inflation will increase too while GDP and real wages will decrease. But this is not the end of the story. Indeed, the second consequence will be a reduction in the price of the imported goods given that now the foreign country is more productive. As a result, CPI inflation in the home country may actually decrease therefore generating an increase in real wages. Finally, since the foreign country is richer, there will be an increase in the demand for home produced goods from foreign consumers and therefore home output will increase too. Therefore, the dynamics of oil price, inflation, real wages and GDP observed in the 2000s may be consistent with an increase in oil demand driven by the increased oil consumption in countries like China and India5. Up to now the theoretical literature (and also most of the empirical one6) has always considered exogenous increases to the oil price driven by reduction in oil supply and has tried to understand whether the oil price crises can be considered responsible for the high inflation and low output of the 70s. The novelty of this

1"Oil prices - New friendships and petropuzzles" Jan 26th 2006, The Economist.

2Spot Price: West Texas Intermediate, Source: Dow Jones & Company, Release: Wall Street Journal. FRED Economic Data. 3See section 3 for a detailed analysis of the data.

4For a detailed review of the literature on oil shocks presenting also alternative hypothesis see section 2.

5Of course, at the same time more then one shock is responsible for the behaviour of the oil price and of the macro variables. Here we are focusing on the shocks we think to have played a major role in the oil price increase in the different years.

(8)

MAGYAR NEMZETI BANK

paper is to develop a theoretical model where the oil price is modelled endogenously therefore, allowing for the study of factors different from oil supply shocks that can affect its dynamics.

To this purpose, a two-country model like the one developed in Claridaet al.(2002) is used, with two main differences. First, in the production sector two inputs are needed, labour and oil. Oil is considered as a traded good therefore, the price is determined internationally and will depend on the demand of both countries (the importance of each country’s oil demand depending on the relative size of the country). In order to simplify the analysis, no one of the two countries is an oil producer. It is assumed that at each point in time there is a world oil endowment which may be subject to exogenous shocks. As in Claridaet al. (2002), firms’ pricing decisions are subject to Calvo staggering. This is important to link the behaviour of real and nominal variables in order to study simultaneously the behaviour of GDP and inflation. Second, in order to have the monetary authority facing a non trivial trade off between inflation and output gap stabilization, the presence of a wage mark-up fluctuating over time is endogenised by allowing for wage rigidities, as in Erceget al.(2000). The reason for this is twofold. First, it allows to study also the dynamic of real wages and wage inflation in a realistic environment. Second, it will make it possible to study how the conduct of monetary policy is going to affect the results in further extensions.

For the sake of simplicity, the main results of the model are derived under the assumption of perfect sym-metry among countries and using a simple interest rate rule targeting national CPI inflation in each country. Two kinds of shocks are considered. The first one is a shock to the oil supply, calibrated such that the real price of oil increases, on impact, by 100%. Such shock is an aggregate disturbance that, given the assumed symmetry, affects both countries in the same way. The second shock is an increase in foreign productivity calibrated in order to match the average quarterly increase in China GDP between 2001 and 2005. Since this shock is meant to represent the increase in productivity in China, a very high autocorrelation coefficient is assumed for this shock, as an approximation for a process that is likely to be stationary only in first differ-ences. The behaviour of real oil price, output, CPI inflation, GDP deflator inflation (from now on domestic inflation), real wages and wage inflation under the two shocks is analysed. The focus is on the behaviour of the variables of the Home country that is supposed to represent the U.S.

After the shock to oil supply the price of oil rises, on impact, by 100%, and then goes back to its steady state level in one year. This is consistent with the experience of the 70s, where the oil price increase was sharp, but not with the current experience, where the real oil price rise is more smooth and prolonged7. Because of the increase in the production costs, the home economy experiences a sharp decrease in output, an increase in price inflation (both CPI inflation and domestic inflation) and wage inflation and a decrease in real wages. The model matches the change in those variables experienced in the U.S. after the 1973-74 shock. It does not match instead the sign of the movements in the same variables after the 2003 shock.

On the contrary, after the shock to foreign country productivity (matching the increase in production in China in the last years), the model predicts an increase in oil price that can account for a 40% increase in the oil price in the period considered. It also generates a reduction of CPI inflation and an increase in real wages, due to the fact that home consumers can now buy foreign produced goods at a cheaper price. Finally, both output and domestic inflation increase. Also for these two variables, the sign is the same as the one experienced by U.S. variables, therefore confirming the original hypothesis that a change in the nature of the shocks driving up oil price is a good candidate to understand the differences between the oil shocks of the 70s and the oil shock of the 2000s. It also tells us that we should not always expect an increase in inflation and a decrease in output every time we observe an increase in the price of energy goods.

Clearly, the model in its current form can not account for the total increase in oil price of the last few years but this is not surprising for at least two reasons: first, we are using an extremely simple version of a two-country model; second, most likely the increase in China’s oil demand is not the only determinant of the current increase in oil price.

7The current increase in oil price started in 2002 and is not yet over i.e., more than a one time shock is a 5 year period of price increase. In this period, the maximum quarter-by-quarter increase has been 18% (second quarter of 2002) while the average quarterly increase has been around 7%. Instead, in the 1973 oil shock, oil price increased by 82% in the first quarter of 1974.

(9)

INTRODUCTION

In order to keep the model as tractable as possible, it has been derived under a set of simplifying assump-tions. Two important ones are the assumptions of perfect risk sharing and perfect symmetry across the two countries, two hypothesis that are clearly unrealistic especially given that the home country is supposed to represent the U.S. while the foreign country represents China. A version of the model with home bias is currently under development.

The paper proceeds as follow: section 2 reviews the related literature, section 3 presents some data evidence to clarify the different behaviour of the U.S. main variables during the different oil shock episodes, section 4 contains the model, section 5 shows the impulse responses of the variables to different types of shocks and section 6 concludes.

(10)

2

Related literature

There have been several papers studying the impact of oil shocks on the economy. Most of the literature has focused on the oil shock episodes of the 70s. An important contribution is the one by Hamilton (1983). He addresses two questions: first, the causes of the oil shocks between 1948 and 1972; second, whether there is a causality relation between the oil price increase and the recession episodes experienced by the U.S. His conclusions are first, that the rise in oil price was driven by either political crises in the Middle East or OPEC decisions, both bringing about a reduction in oil supply therefore pushing up the oil price. Second, those shocks appear to be important in explaining the periods of recession and high inflation experienced by the U.S. economy. For a different view on the relation between oil price and macroeconomy after 1973 see Hooker (1996) (and the reply by Hamilton (1996)) and Barsky & Kilian (2004)8.

Bernankeet al. (1997) and Barsky & Kilian (2002) argue that the conduct of monetary policy is crucial in explaining the periods of recessions and high inflation occurred after the oil shocks of the 70s.

On the theoretical side, Rotemberg & Woodford (1996) and Finn (2000) show that oil price shocks have a substantial impact on economic activity despite the fact that the proportion of oil used in the production process is relatively small. Both papers are in closed economy and with no nominal rigidities. Also, since they are interested in supply shocks, they use an exogenous process for the oil price. The paper by Backus & Crucini (2000) is a three-country model with no nominal rigidities, that shows that oil price shocks account for a big proportion of terms of trade volatility. They endogenise the oil price through the presence of a third country producing oil. Since there are no nominal rigidities this framework is not suitable for monetary policy analysis. Another work is the one by Leduc & Sill (2004). It is again a closed economy with an exogenous process for the oil price, but it is the first one with nominal rigidities. The main objective of the paper is to see whether the recessionary consequences of an oil shock are only due to the oil shock itself, or also to how monetary policy is conducted. They use a DSGE model to show that monetary policy plays only a secondary role in the recessionary process, but that a monetary authority more concerned about inflation better deals with the problem. Also in this case, the focus is on supply shocks, with the price of oil modelled as an exogenous process.

Two papers more closely related to the present one are Blanchard & Galí (2007) and Kilian (2007c). Like the present paper, Kilian (2007c) underlines the importance of identifying supply vs demand shocks to oil price. He provides a decomposition of real oil price into: oil supply shocks; shocks to the aggregate global demand for industrial commodities; demand shocks that are specific to the oil market (i.e., precautionary oil demand). Using this decomposition he claims that, while the oil price increase in the 70s is mainly do to precautionary demand increase, in the current increase a crucial role is played by aggregate demand shocks. However, to be notice is that the identified aggregate demand shocks "rise U.S. real GDP growth in the first year after the shock, but lower it in the second year. They also cause a persistent increase in CPI inflation9". This is not really consistent with the current U.S. situation where we do not observe an increase in CPI (see section 3). Blanchard & Galí (2007) try to understand the difference between the various oil shocks but their focus (differently from the present paper) is on the fact that the oil shocks seem to have a far lower impact on economic activity now than in the 70s10(in line with the literature on the great moderation) rather then on understanding the different directions in the movements of output and CPI following an oil price increase. They start from the assumption that the source of the change in oil price is always the same, i.e. an exogenous increase in oil price, and study how a different environment can affect the transmission of the same shock. As possible candidates, they consider differences in the monetary policy, in the degree of wage rigidity and in the proportion of oil used in the production and show that a change in each of them can reduce the volatility of both prices and quantities in response to the same oil shock. However, given that they always focus on

8See also Kilian (2007a) and Kilian (2007b) 9Kilian (2007c), page 3.

(11)

RELATED LITERATURE

supply shocks, their model is not suited to understand how an oil price increase can be accompanied by an increase in output and a reduction in CPI like it happened in the U.S. in 2000s.

To be noticed is that the story presented in Blanchard & Galí (2007) and the one developed in this paper are complements rather than substitutes. Indeed, it is reasonable to expect that shocks of different nature are hitting the economy at the same time, and also that the structure of the economy is evolving over time. The current increase in oil price is likely to be the consequence of both supply and demand shocks. The kind of changes in the structure of the economy analysed by Blanchard & Galí (2007) reduce the reaction of inflation and GDP to the oil price increase, i.e. reduce the increase in inflation and the decline in output following the oil price increase. If at the same time also a demand shock boosting oil price is at work, we show in the present paper that inflation decreases and output increases. The two shocks together increases oil price but offset each other in terms of movements in inflation and output and this explain the reduction in the volatility of those variables. Finally, if the demand shock is sufficiently strong, the overall result can be a positive growth in GDP and low or decreasing inflation, as observed in the U.S. in 2000s.

(12)

3

Stylized facts

Figure11(1) reports the behaviour of oil price from the first quarter of 1960 to the first quarter of 200812. The top panel represents the nominal oil price expressed in dollars per barrel. In the bottom panel the nominal oil price has been divided by the CPI in order to express it in real terms. Taking logs it is possible to interpret the differences between two periods as percentage changes. The 1960 is the base year.

Following Blanchard & Galí (2007), we define an oil shock as an episode where the cumulative increase in oil price has been of more then 50% and has lasted for more than one year. Following this criteria four oil shocks are identified and for each of them the date at which the 50% threshold is reached has been reported in the graph. For each oil shock, table 1 reports the overall increase in the real oil price. In terms of the magnitude of the oil price increase, the four episodes are all alike, with an increase of oil price of around 100%. The main difference is that, while in the first two episodes (and, too a lesser extend, also in the third one) few quarters were needed to reach the 100% increase, in the last one the increase is much smoother. Therefore, the dynamics of the change in oil price in the last episode seems to differ from the first two.

The next step, after having identified the shock episodes, is to look at what happened to other variables during those periods. In pictures (2), (3) and (4) the following U.S. variables are represented: CPI inflation, GDP deflator inflation, real GDP growth rate, real wage13 and wage inflation. The inflation rates and the growth rate are annualized rates while the real wage is expressed in logs with the 1960 as base year.

The first two episodes of oil price increase coincide with an increase in all inflation variables and a decrease in GDP growth and real wage. On the contrary, the last episode coincides with a positive GDP growth rate, an increasing real wage and low inflation rates. Things are even more clear looking at the results of tables 2 and 3. Following the same methodology used by Blanchard & Galí (2007), for each variable, table 2 (3) presents the percentage change between the average value one (two) year(s) after the shock and the average value one (two) year(s) before the shock. As reference quarter it has been used the one where the oil price reached the 50% increased needed to qualify the increase as an oil shock. Therefore, the interpretation of the table 2 is that average CPI inflation in the 1974 (the year after the shock) was 3.2 percentage points higher then it was the year before. At the same time the real GDP growth was 6 percentage points lower than in the previous year. During the first two oil shocks the U.S. economy experienced a period of stagflation with rising inflation and decreasing output. At the same time the real wage also decreased after the oil shocks while wage inflation was rasing. Results are basically unchanged if you consider a longer horizon (table 3). Things are drastically different in the last two shocks and in particular in the one of 2003. During this last oil price increase, the U.S. economy experienced a positive GDP growth, an increase in real wages, a decrease in wage inflation and either a decrease (at the 4 quarters horizon) or only a moderate increase (at 8 quarters horizon) of CPI inflation.

Those simple observations underline how much different is the current experience from the oil shocks of the 70s.

Let us now look at the evolution of oil consumption over time. Figure (5) represents the average yearly world consumption (expressed as thousand barrels per day), with the relative contribution of OECD and non-OECD countries reported. As for the previous pictures, the oil price shock episodes are also reported. While, after the oil shocks of the 70s, world oil consumption decreased, this is not the case in the 2000s where, even with an increase of the oil price of more then 100%, the world oil consumption has increased steadily. This of course is not surprisingly given that in the 70s several countries experienced a recession while this is not the case today. Still, after a supply shock we should not expect to observe an increase in oil consumption. Figure (6) reports the evolution of crude oil production for the period 1970-2007 and shows no contraction in oil production when real oil price reached a cumulative increase of more than 50% in 2003.

11Tables and figures follow at the end of the paper.

12See appendix for a detailed description of the data used in the paper 13Computed dividing the nominal hourly earnings by the CPI.

(13)

STYLIZED FACTS

Table 4 and 5 are useful to understand the role played by different countries in the dynamics of oil consump-tion. In particular, table 4 reports the growth rate of oil consumption for the period 2001-2006 for different areas in the world. Two things are worth noticing. First, most of the oil consumption growth seems to be driven by non-OECD countries which experienced a growth rate of around 20% against the 3% only of OECD countries. Second, China with its 46% is the country with the highest growth rate in that period14. Table 5 reports the contribution to the world oil consumption for the year 2006 of each geographical area reported in table 4. The contributions of the single countries are also reported when they account for at least around 2% of world oil consumption. Several things are worth noticing. Today non-OECD countries account for more than 40% of world oil consumption (while back in the 70s they accounted for only 26%). In particular, China’s share of world oil consumption (8.51% of the world consumption) is second only to that of the United States (24.45%) and is far above the one of any single other country15. Being the coun-try with the second biggest share of oil consumption and having experienced the highest growth rate in oil consumption during the 2000s, it seems worthy trying to investigate the role played by the dynamics of this country in shaping the behaviour of oil price nowadays.

Finally, since the model developed in the next section is an open economy one and trade plays a crucial role, it is worthwhile to investigate the dynamics of U.S. imports and exports in the last 30 years. Figures (7) and (8) report the yearly average of U.S. imports and exports by country of origin and destination16for the period 1974-2007. While imports seem to have grown at a higher rate starting from 1995, China seems to be the country who benefitted the most from this tendency. Starting from 2001 imports from China have grown at a much higher rate than the one from the other trade partners. As a consequence of that, in 2007 China overtook Canada becoming the country exporting the most in the U.S. market. A similar tendency can be seen in the export market (figure (8)) even though the speed here seems to be somehow lower. Canada and Mexico are still the main destination countries for the U.S. exports, but starting from 2001 we can observe a sharp increase in the exports towards China which already overtook U.K, Germany, France and Japan becoming the third destination country for the U.S. exports.

To sum up, while the oil shocks of the 1970s are associated with increasing inflation (both in prices and wages), decreasing real GDP and decreasing real wage, the opposite is true for the current increase in the oil price where the US economy in the period 2000-2006 is characterized by stable inflation, increasing real wage and positive real GDP growth. At the same time, while oil consumption decreased during the first oil price shocks, this is not the case in the 2000s. Particularly interesting seems to be the case of China which experienced a vary high growth rate becoming the second country in terms of oil consumption in the 2000s. During the same period in U.S. both imports from and exports to China increased considerably.

In the next section a model is provided that reconciles those different evidences and shows the role which imports and exports may have played in contrasting the effects of oil price increases on U.S. economy in the 2000s, therefore providing an explanation of why the 2000s look different from the 70s.

14There are a few countries with an even higher growth rate: Moldova, Qatar, Benin, Cape Verde, Mozambique, Seychelles, Togo, Fiji, Macau, Maldives, Papua New Guinea and Vietnam. However, they account all together for only 0.6% of the world oil consumption so they are negligible.

15As you can see from table 5 Japan is also a "big player" accounting for 6.10% of world oil consumption. But during that period Japan was experiencing very low (sometimes negative) real GDP growth rates and oil consumption in Japan during the period 2001-2006 decreased by 4.35%.

(14)

4

The model

The world is populated by a continuum[0, 1]of agents. Agent h ∈[0,n]lives in the home country (H) while agent h∈(n, 1]lives in the foreign country (F). Therefore, the two countries may have different size. For everything else, perfect symmetry is assumed. Variables are expressed in per capita terms. The reference model is Claridaet al.(2002). We follow their modelling strategy assuming that in each country there are as many final goods producers as households. In the final goods sector perfect competition is assumed. In order to keep the model as simple as possible, it is assumed that in each period there is a world endowment of oil. The world oil price is determined in equilibrium given the oil demand from firms in the two countries. The profits from selling oil are redistributed in each period among all the households of the two countries evenly (i.e. the per-capita share of profits is the same for home and foreign households), as a lump sum transfer.

HOUSEHOLD PROBLEM

Each household supplies a differentiated labour service to each of the firms in the country. As usual in this class of models, labour is immobile across countries. The elasticity of substitution across workers isθw. Since each household acts as a monopolist in the supply of his labour, he chooses the wage in order to maximize the lifetime utility, subject to the labour demand schedule.

Householdhin the domestic economy chooses{Ct,WH,t(h),Dt+1}in order to maximize:

E0    ∞ X t=0 βt    Ct1−σ 1−σ − Nt(h)1+ϕ 1+ϕ       (1) subject to: PtCt+Et[Qt,t+1Dt+1] = (1+τw)WH,t(h)Nt(h) +Dt−Tt+ Γt+ Υt (2) Nt(h) =   WH,t(h) WH,t   −θw Nt (3) with: Ct=CHn,tCF1−,tn Pt=κ−1PHn,tPF1−,tn WH,t= 1 n Z n 0 WH,t(h)1−θwd h 1−θw1 (4) where Qt,t+1 is the stochastic discount factor, Dt is the payoff in t of a portfolio held in t −1 and κ

nn(1−n)1−n. Equation (3) is the labour demand, obtained solving the cost minimization problem of the firms. τw is a subsidy to labour that can be used by the fiscal authority in order to offset the distortion created by the presence of monopolistic competition in the labour market. Γt and Tt are two lump-sum components of household income representing, respectively, dividends from ownership of firms and taxes. Υt is the lump-sum transfer from the redistribution of profits from the sale of oil. Note the Cobb-Douglas aggregator for consumption implies the following assumptions: unit elasticity of substitution between home and foreign produced goods; no home bias in consumption; the number of final producer firms coincide with the number of households in each country. Far from being realistic, those assumptions are imposed here only for the sake of simplicity.

(15)

THE MODEL

Consumption decision and intertemporal optimization From the expenditure minimization problem we obtain:

PH,tCH,t=nPtCt PF,tCF,t= (1−n)PtCt (5) while, combining the first order conditions with respect toCtandDt+1, we find the standard Euler Equation:

1=βRtEt   ‚C t+1 Ct Œ−σ P t Pt+1   (6) withRt =Et(Q1t,t+1). Wage decision

Following Erceg et al. (2000), it is assumed that in each period only a fraction 1−ξw of households can reset wages optimally, while for the othersWH,t(h) =WH,t1(h). Therefore, each household maximizes its lifetime utility taking into consideration that, with probabilityξwT, in periodT his wage will still beWH,t(h). Given this, the first order condition with respect to wage is:

Et ∞ X T=0 (βξw)T – Ct−σ+TWH,t(h) Pt+T (1−Φw)−Nt+T(h) ϕ™N t+T(h) =0 (7) with 1−Φw = (1+τw)θw −1

θw . WhenΦw=0 the fiscal authority completely offset the distortion caused by monopolistic competition in the labour market.

When wages are flexible,T =0 and equation (7) becomes: WH,t(h) Pt =Nt(h) ϕCσ t 1 1−Φw (8) andWH,t(h) =WH,t andNH,t(h) =NH,t ∀h∈[0,n].

Under sticky wages we need to log-linearize equation (7), obtaining the following wage inflation equation: πw,t=−λwµbw,t+βEt[πw,t+1] (9)

whereµbw,t =log(WH,t)−log(Pt)−ϕlog(Nt)−σlog(Ct) +log(1−Φw), πw,t is the log of wage inflation,

andλw=1−ξw

ξw 1−βξw 1+ϕθw.

International tradability of state-contingent securities

Because of international tradability of state-contingent securities, the intertemporal condition for foreign consumers can be written as:

β   Ct+1 Ct∗   −σ P∗ t Pt+1 et et+1 =Qt,t+1 (10) whereet is the nominal exchange rate. Since there is no home bias in consumption and we assume that the law of one price holds, i.e.Pt=Pt∗et, thenCt=Ct∗ ∀t.

(16)

MAGYAR NEMZETI BANK

FIRM PROBLEM - FINAL GOODS SECTOR

Intermediate home produced goods are aggregated into final goods using the following technology:

Yt=   Z1 0 Yt(j) θp−1 θp d j   θp θp−1 (11) In both countries there is a continuum [0, 1]of producers in the intermediate sector. In each country the final goods are produced using only intermediate goods produced within the country. There is a continuum [0,n]of final good producers in the home country and a continuum(n, 1]in the foreign country. The final goods sector operates in perfect competition. Profit maximization yields to:

Yt(j) =   PH,t(j) PH,t   −θp Yt (12) withPH,t =”R01PH,t(j)1−θpd j— 1 1−θp.

FIRM PROBLEM - INTERMEDIATE GOODS SECTOR

Intermediate goods sector firms produce accordingly to the following technology:

Yt(j) =AtNt(j)αOH,t(j)1−α (13)

whereOH,t(j)is the oil demand of firm "j",Nt(j) =

1 n Rn 0 Nt j(h) θw−1 θw θwθw −1

is firm "j" labour demand and the technology process is defined as:

at+1=ρaat+"A,t. (14) withat≡log(At)and where"A,t is an i.i.d shock with zero mean. Wheneverα=1 (i.e. oil is not used in the production function) we are back to the standard case. Each firm has to choose how to optimally combine labour and oil and, also, how much to demand of each labour type. Solving the cost minimization problem leads to: WH,t Po,t = α 1−α OH,t(j) Nt(j) (15) Nt j(h) =   WH,t(h) WH,t   −θw Nt(j) (16)

where Po,t is the price of oil (it is determined endogenously when computing the equilibrium in the oil market). Integrating (16) over all firms, we obtain the labour demand schedule for workerh, (3).

(17)

THE MODEL

Marginal cost

Using (15) we can derive the following expression for firm nominal marginal cost: M CH,t(j) = 1 α α 1−α 1−α 1 AtW α H,tP 1−α o,t (17)

As standard in this class of models, the marginal cost is not firm specific. Whenα=1 the expression for the marginal cost simplifies to the usualM CH,t=WH,t

At .

Pricing decisions

Firm j production is small with respect to the world production and the same is true with respect to firm j oil demand. Therefore, when undertaking production decisions, firm j takes oil price as given. This means that pricing decision is isomorphic to the one we obtain in the standard case. This implies that, assuming Calvo price setting and beingξp the probability of not being able to reoptimize next period, if firm j is allowed to reoptimize int, it will choosePH,t(j)such that:

Et ∞ X T=0 ξT p Qt,t+TYt+T(j)  PH,t(j)− 1 1−Φp M CH,t+T  =0 (18) with 1−Φp= (1+τp) θp−1

θp . WhenΦp=0, the fiscal authority is completely offsetting the distortion coming from the presence of monopolistically competitive goods market.

Under flexible prices equation (18) simplifies to: PH,t= 1

1−Φp

M CH,t (19)

i.e. the price is a constant mark-up over the marginal cost.

To solve the model under sticky prices, we need to log-linearize equation (18) around its the steady state, obtaining the Phillips Curve on home inflation:

πH,t=βEt[πH,t+1] +λmcÓt (20)

whereπH,t=log PH,t PH,t−1,λ

(1−ξp)(1−βξp)

ξp , andmcÓt=logM CH,t−logPH,t−log(1−Φp)is the log-deviation of the real marginal cost from the flexible price allocation.

EQUILIBRIUM CONDITIONS

Oil market equilibrium

Recall that up to now all the variables have been expressed in per-capita terms. As a consequence, to compute the world oil demand we need to multiply the per-capita oil demand coming from each country by the country size. The oil demand of the world economy is therefore:

Otd≡n Z1 0 OH,t(j)d j+ (1−n) Z1 0 OF,t(j)d j (21)

(18)

MAGYAR NEMZETI BANK

Since the focus is not on the consequences for an oil producer country of an increase in the oil demand, by assumption none of the countries is producing oil. To simplify the model as much as possible, it is assumed that at each point in time there is a world oil endowmentOts. We already clarified when studying the household’s optimization problem that profits from selling oilPo,t∗Otsare evenly redistributed as lump sum transfer to the world consumers. The reason for this is twofold: first, we do not want to deal with a third country producing oil and second, we do not want to introduce a source of asymmetry across the two countries, this is why the amount of per-capita profits redistributed is the same across the two countries. To account for supply shocks to the oil price, an exogenous, i.i.d. shock ξt to the otherwise constant oil endowment is introduced. In order not to complicate the analysis too much, oil is assumed to be non storable i.e., oil supplied in periodtis consumed in the same period. The world oil supply is defined by the following process:

Ots+1= (Ots)ρoeξt+1 (22) where,ξt is an i.i.d. exogenous shock to the supply of oil. Whenξt=0∀t,Ot=O=1∀t.

The total oil demand of home country is: OHd,t≡n Z1 0 OH,t(j)d j =n1−α α WH,t Po,t Nt (23)

The total oil demand of foreign country is17:

OFd,t≡(1−n) Z1 0 OF,t(j)d j= (1−n)1−α α W∗ F,t Po,t N ∗ t (24)

For the oil market to be in equilibriumPo,t must verify18: Po,t Pt = 1 Ots 1−α α  n WH,t Pt Nt+ (1−n) W∗ F,t Pt∗ N ∗ t   (25)

An exogenous shock to oil supply will affect the equilibrium price of oil throughξt. An exogenous shock to the productivity of one of the two countries will change the quantity produced by both and, therefore, it will affect the oil price through a change in the oil demand. Clearly, the bigger the country, the bigger the consequences of a change in his production.

Asset market equilibrium

Under the two assumptions that asset markets are complete and that the low of one price holds, the equilib-rium condition implies thatCt=Ct∗ ∀t.

Equilibrium in the goods market

Like in CGG, goods market clearing conditions imply that:

PH,tYt=PtCt PF,tYt∗=Pt∗Ct∗ (26) 17Real variables with a star refer to the foreign country. For nominal variables, the subscript F refers to the foreign country and a star means that they are expressed in the foreign currency.

18Because of the low of one price, Pt Po,t =

P∗

t P∗o,t.

(19)

THE MODEL

Therefore,

Ct=κYtn(Yt∗)1−n (27) where the aggregate output in equilibrium is:

Yt= AtN α t O 1−α H,t Zt (28) withZt=R1 0 P H,t(j) PH,t −θp d j.

Now that we have all the first order and the market clearing conditions, it is possible to study the dynamics of the model both in the case of no nominal rigidities (section 4) and under sticky prices and wages (section 4).

FLEXIBLE PRICES AND WAGES

Let us first study the case in which prices and wages are perfectly flexible in both countries. Using equations (8), (15), (19), (26), (27) and (28)19, it is possible to derive the following expression for the natural level of output of the home country:

Yt=αA1 1α α (1−α)A2 κA3[1 −Φp]A1[1−Φw]A4A A2 t    Po,t Pt    −(1−α)A2 (Y∗t)(1−n)A3 (29) where: A1= 1+ϕ(1−α) 1+ϕ−n[1+ϕ(1−α)ασ] A2= (1+ϕ) 1+ϕ−n[1+ϕ(1−α)ασ] A3= 1+ϕ(1−α)ασ 1+ϕ−n[1+ϕ(1−α)ασ] A4= α 1+ϕ−n[1+ϕ(1−α)ασ]

Doing the same for the foreign country and using the clearing condition for the oil market (25), the equilib-rium oil price must satisfies:

Po,t Pt =   1 Ots   1+ϕ(1−α) α 1α α – 1 1−Φw ™ ” κYnt(Y∗)1−n—σ n”YtA−t1— 1+ϕ 1+ϕ(1−α)+ (1n)” Y∗t(A∗t)−1— 1+ϕ 1+ϕ(1−α) 1+ϕ(1−α) α (30) 19In the flexible price equilibriumZ

(20)

MAGYAR NEMZETI BANK

Taking a log-linear approximation around the symmetric steady state of (29), of its counterpart for the foreign country and of (30) for the special case of log utility in consumption (i.e. σ=1), it is possible to rewrite the natural level of output in the two countries and the real oil price as function of only exogenous shocks20:

ˆ yt=at+ (1−α)ξt (31) ˆ y∗t =a∗t+ (1−α)ξt (32) ÔPo,t Pt =nat+ (1−n)a ∗ t−αξt (33)

The following things are worth noticing. First, the natural level of output does not move in reaction to foreign technology shocks. This is a standard result under the assumption of log utility in consumption that goes through even if oil (a tradable good) is introduced in the production function. Second, the natural level of output in the two countries reacts to global oil supply shocks21. The sensitivity of the natural level of output to those shocks depends crucially on the proportion of oil used in the production function, 1−α. Third, the oil price reacts both to exogenous oil supply shocks and to technology shocks in the two countries. The size of the country is crucial in determining the impact of its own technology shock on the global price of oil. Finally, under flexible prices and wages and when perfect symmetry across the two countries is assumed, the way a technology shock is transmitted to the other real variables does not depend onαi.e. the dynamics are the same that would arise under the standardYt=AtNt. In particular, if we consider the case of a home technology shock (settingξt=a∗

t =0) it is possible to show the following: ˆ yt=at yˆ∗t =0 ˆ ct = ˆct∗=natt= ˆn∗t =0 ˆ wt= ˆw∗t =nat ˆoH,t= ˆoF,t=0

In the next section we analyze the transmission mechanism when both prices and wages are sticky.

STICKY PRICES AND WAGES

In this section we come back to the case where prices and wages are sticky in both countries. Using the log-linearized version of the equilibrium conditions in section 4 and of the first order conditions derived in sections 4, 4 and 4, we can derive the New Keynesian Phillips Curve (NKPC) from equation (18)22:

πH,t=βEt[πH,t+1] +λ ” Aµbw,t+B(yt−yt) +Cµb ∗ w,t+D(y ∗ t −y ∗ t) — (34) where A≡ α+ (1−α)n(1+ϕ) 1+ϕ(1−α) B≡ – σn+1−n+ϕ α 1+ϕ(1−α)+n(1+ϕ) 1+ϕ(1−α)α α[1+ϕ(1−α)] ™ C ≡ (1+ϕ)(1−α)(1−n) 1+ϕ(1−α) D≡ (1−α)(1+ϕ)2(1−n) α[1+ϕ(1−α)] (35) 20As standard,x

t=l o g(Xt)whileˆxtrepresents log deviations from the steady state. 21Recall that a positiveξ

tmeans an exogenous increase in oil supply. It has an expansionary effect on output in both country. 22For the complete derivation see Appendix A

(21)

THE MODEL

and yt and y∗t represent, respectively, the log of the level of output in absence of nominal rigidities in the home and foreign country. µb

w,t represents the log-deviation of the wage mark-up from its frictionless level, in the foreign country.

It is useful to study how the NKPC becomes in some special cases. No oil used in production, i.e.

α

=

1

In this case,C =D =0,A=1 andB =σn+1−n+ϕ, i.e. we are back to a standard NKPC with sticky wages. Because of the presence of sticky wages, the monetary authority faces a trade of between stabilizing output gap and stabilizing inflation.

General case

Whenα <1, i.e. when oil is used in the production function, the Phillips Curve of the home country is function also of the foreign output gap and of the foreign wage mark-up fluctuations. Therefore, the new assumption on the production process has amplified theopen economy dimensionof the model. The reason for this is simple. Even if, when solving their optimization decisions, foreign producers take home variables and oil price as exogenously given, in equilibrium their production decisions affect oil price through their impact on world oil demand. Therefore, through their impact on world oil price, foreign producers influence home marginal costs. Specifically, when foreign output is above its natural level, foreign oil demand is also going to be above the optimal level and this creates an upward pressure on oil price. Also, when the average wage markup charged by foreign workers is above the one charged under flexible wages, there is going to be a partial substitution between labour and oil in the production process and this will also generate an upward pressure on the oil price.

How much close we are to the standard case depends both onαand n. The role played byαis clear. As αapproaches to 1, the role of oil in the production process approaches to 0, therefore we are closing this channel and going back to the standard model. On the other hand, asn approaches 1, the foreign country becomes small and, therefore, it is unable to affect the oil price that is determined at the world level. As a consequence, no matter how big can be the role played by oil in the production process (i.e. no matter how muchαcan be close to zero), since the oil price will not be affected in a significant way by foreign variables, this new channel will not play an important role andC andD will approach to 0.

In the next section we will use impulse response functions to analyse the different transmission mechanism implied by a supply versus a demand shock.

(22)

5

Demand vs Supply Shock: Impulse Response

Analysis

BASELINE CALIBRATION

The two countries are assumed to be perfectly symmetric. They also have the same size, i.e. n=0.5. Most of the parameters have been set following Galí & Monacelli (2005).

Consumer

The discount factorβis set equal to 0.99, implying a riskless annual return of around 4%. The coefficient of relative risk aversionσ is set to 1, implying log utility in consumption. ϕ=3 so that the labour supply elasticity is 1/3. θw = 6 implies a wage markup of 1.2. Finally, ξw = 0.75 i.e. nominal wages adjust, on average, once a year.

Firm

Like for workers,θp=0.6 andξp=0.75. The share of oil in the productionαis set equal to 0.05. Fiscal subsidies

The fiscal authority sets the subsidies τw and τp such that the flexible (prices and wages) equilibrium is efficient from the single country point of view.

Monetary Policy

It is assumed that the monetary authorities in both countries follow an interest rate rule targeting CPI infla-tion with a coefficient of 1.5.

Exogenous Shocks

Two kind of shocks are considered. A negative shock to oil supply (exemplifying the oil shocks of the 70s) and a positive shock to foreign productivity (aimed at capturing a demand shock). The foreign productivity process is assumed to be very persistent with ρA = 0.999. On the contrary oil supply is assumed to be much less persistent withρo=0.5. When studying the impulse responses to a foreign productivity shock we assumec o r rA,A∗=0 in order to be able to disentangle the impact of foreign pressure on oil demand alone.

DEMAND VS SUPPLY SHOCK

Two experiments are conducted. The first one is to simulate the model under an oil supply shock that generates an increase of 100% in the real price of oil and that dies off very fast (the correlation coefficient of the process for the oil supply is indeedρo =0.5). Since the increase in the 70s of the oil price was always happening in a few quarters, this experiment is meant to replicate such circumstances. Figure 9 reports the impulse responses of the real price of oil plus the following home variables: CPI inflation, real output, domestic inflation, real wage and wage inflation. This shock reproduces in the model an episode of stagflation

(23)

DEMAND VS SUPPLY SHOCK: IMPULSE RESPONSE ANALYSIS

like the one observed in the 70s with a contraction in output accompanied by a rise in inflation (measured both in terms of domestic inflation and CPI inflation), a rise in wage inflation and a contraction of real wages. This shock is symmetric for both countries therefore the impulse response functions for the foreign country are exactly the same as for the home country. The reason behind those dynamics is that such a shock increases production costs in both countries, therefore it fuels inflation, both in terms of domestically produced goods and with respect to goods produced abroad. This decreases real wages and contracts total output. More in detail, the supply shock generates a reduction in the home output of 0.9%, consistent with a reduction in the yearly growth rate of GDP of around 3.6%. Inflation (measured both in terms of CPI inflation and domestic inflation) increases by 0.6%, implying an increase in the yearly rate of 2.4%. This generates a decrease in real wages of 0.4% i.e., a reduction in one year of 1.6% accompanied by an yearly increase in wage inflation of around 0.6%. The direction in the changes of those variables is consistent with the one observed in the 70s even if the magnitude is lower than the one reported in table 2. It is important to note however that, given that the supply side of the oil market is not explicitly modelled, while conducting this experiment it is not possible to evaluate whether the quantity reduction in the oil supply needed to deliver the 100% increase in the oil price is consistent with what happened in reality. This problem is common to all papers mentioned in the literature review which studies the transmission of oil price shocks considering an exogenous process for the oil price. In this model a more careful experiment can instead be conducted when studying a demand shock.

In the second experiment the idea is to approximate the increase in demand driven by the growth process involving Asian countries. As a first approximation, we calibrate the shock to the foreign technology in order to deliver an increase in foreign output (on impact) of 3%. This is approximately consistent with the quarterly growth rate experienced by China between 2001 and 2006. The results are shown in figure 10. We are aware the the correct experiment would be to consider an increase in the growth rate of the foreign country. As a first approximation, we set the autocorrelation coefficient for the technology process close to one.

Since the foreign country is more productive, foreign produced goods become cheaper. Therefore, there will be an increase in the demand for those goods from foreign consumers (that are now richer) but also from home consumers. At the same time also the demand for home produced goods will increase because of the income effect affecting foreign consumers. As a consequence, home output is going to increase. The increase in production in both countries drives up world oil demand and, therefore, the oil price. This increases domestic inflation in the home country because of the increase in the production costs (the home country is producing more without being more productive). Home CPI inflation instead decreases because now home consumers are importing goods from abroad at a cheaper price23. This explain how we can be experiencing a reduction in CPI inflation and an increase in domestic inflation and output together with an increase in the oil price. Real wages increases because of the reduction in CPI. The direction of the change in those variables is consistent with the one experienced by the U.S. variables in the 2000s (see table 2). The only difference between the model and the data is in the behaviour of wage inflation. While wage inflation is decreasing in the data, it is increasing in the model. Two things are worth noticing. The first one is that even if wage inflation is increasing under the demand shock, this increase is lower then the one experienced under a supply shock. The second is that the direction of the change in wage inflation depends very much on the persistence of the technology shock. WhenρA=0.9, wage inflation decrease after a positive shock to foreign technology. Quantitatively, the shock delivers an increase in the real oil price of 1.5% in one quarter. Since China experienced this growth rate for around 24 quarters, it can explain an increase in oil price of around 40%. This is a reasonable results given that other shocks are affecting the world economy at the same time. The movements in the other variables are of the right sign but the magnitude is bigger than the one observed in the data, especially for what concerns the home output. One possible explanation for this is that the model has been solved under the assumption of no home bias. Removing this assumption would capture the fact that most of China consumption is still on home produced goods. This would decrease the impact of the 23Note that this difference in the behaviour of CPI inflation and domestic inflation is consistent with the one experienced by the U.S. in correspondence of the last oil price increase (see table 2).

(24)

MAGYAR NEMZETI BANK

foreign productivity shock on home output. It would also allow to study how much the dynamics of imports and exports in the model do match the data for the U.S. and China. As last remark it may be worth noticing that, under the present specification of the model, considering the case in which the monetary policy in the foreign country keeps the nominal exchange rate constant does not affect the results.

(25)

6

Conclusions

The paper started asking the question of how is it possible that, despite the same magnitude in the increase in the price of oil, the 2000s appear to be so different from the 70s, with a positive output growth, a low (even decreasing) CPI inflation and increasing real wages. An explanation based on different shocks hitting the economy in the two periods, relating in particular the current behaviour of U.S. variables and oil price with the growth in China has been provided. In particular, using a two-country model where the price of oil is determined endogenously, two kind of shocks have been considered: a negative shock to oil supply and a positive shock to foreign country productivity. The model, despite its simplicity, is able to generate changes in the relevant variables under the two types of shocks that are consistent with the ones observed in the data.

(26)

MAGYAR NEMZETI BANK

Appendices

A

Derivation of ˆ

mc

All lower case letters indicate the log of the variables. For a generic variable xt, bxt represents log deviation

from the steady state when there are nominal rigidities, andext stands for log deviation from the steady state

when there are no nominal rigidities. xt characterizes the frictionless level of the variables. For simplicity, either both countries faces nominal rigidities or they both operate in a flexible environment. Writing (17) in log-deviation from the steady state, remembering thatwh,t−pt=µw,t+ϕnt+σct24, and using (26) together with the fact thatbxt−ext=xt−xt, we can write:

Ó mct =α[µbw,t+ϕ(nt−nt)] + (ασ−1)(ct−ct)+ + (1−α)   log ‚P o,t Pt Œ −Po,t Pt   +yt−yt (36)

Taking a first order approximation of (28) around the steady state and using (15), we have:

b

nt=byt−at−(1−α)[µbw,t+ϕnbt−σbct] + (1−α) ÔPo,t

Pt (37)

Taking a first order approximation of (25), using the fact that ct = ct∗, considering the fact that the steady state is symmetric across countries, and that also for the foreign countrywh,t− p∗t =µw,t+ϕn∗t +σct∗, we have: ÔPo,t Pt =−ξbt+nµbw,t+ (1−n)µb ∗ w,t+n(1+ϕ)nbt+ (1−n)(1+ϕ)nb ∗ t+σbct (38)

From (27) we have that:

b

ct =nbyt+ (1−n)by

t (39)

We can write an expression analogous to (37) for the foreign country. Using that equation together with (37), (38) and (39) we can writenbt and

ÓPo,t

Pt as function only of the exogenous shocks (at,a∗t andξt),byt,byt∗,µbtand b µ∗ t: ÔPo,t Pt =− 1+ϕ(1−α) α ξbt− n(1+ϕ) α at− (1−n)(1+ϕ) α a∗t+ + – n(1+ϕ) α +σn ™ b yt+ –( 1−n)(1+ϕ) α +σ(1−n) ™ b y∗t +nµbw,t+ (1−n)µb ∗ w,t (40) b nt=− 1−α α ξt+ α+n(1−α)(1+ϕ) α[1+ϕ(1−α)] (ybt−at)+ +(1−α)(1−n)(1+ϕ) α[1+ϕ(1−α)] (yb ∗ t −a ∗ t) + (1−n)(1−α) 1+ϕ(1−α) (µb ∗ w,t−µbw,t) (41) 24µ

(27)

APPENDIX A

Using the fact thatbxt−ext=xt−xt, we obtain:

Po,t Pt − Po,t Pt = – n(1+ϕ) α +σn ™ (yt−yt)+ + –( 1−n)(1+ϕ) α +σ(1−n) ™ (yt∗−y∗t) +nµbw,t+ (1−n)µb ∗ w,t (42) nt−nt =α+n(1−α)(1+ϕ) α[1+ϕ(1−α)] (yt−yt)+ +(1−α)(1−n)(1+ϕ) α[1+ϕ(1−α)] (y ∗ t −y ∗ t) + (1−n)(1−α) 1+ϕ(1−α) (µb ∗ w,t−µbw,t) (43) ct−ct =n(yt−yt) + (1−n)(yt∗−y∗t) (44) Substituting equations (42), (43) and (44) into (36) we obtain the expression for the log deviation of the real marginal cost that, once substitute into (20), gives equation (34) in the text.

(28)

MAGYAR NEMZETI BANK

B

Data description

All variables refer to U.S.A. All data are available for the period Q1:1960-Q1:2008 with the exception of oil consumption which is available at yearly frequency for the period 1970-2006, oil production which is available at yearly frequency for the period 1970-2007, and the imports/exports series which are available starting from 1974.

Nominal Oil PriceSpot Oil Price, West Texas Intermediate, dollars per barrel, quarterly observations con-structed from monthly data.

Source:http://www.eia.doe.gov/.

Nominal GDPGross Domestic Product, seasonally adjusted annual rate, billions of dollars, quarterly ob-servations. Source: FRED, http://research.stlouisfed.org/fred2/.

GDP DeflatorGross Domestic Product Implicit Price Deflator, seasonally adjusted, quarterly observations, index 2000=100. Source: FRED, http://research.stlouisfed.org/fred2/.

Consumer Price IndexConsumer Price Index For All Urban Consumers: All Items, seasonally adjusted,

quarterly observations constructed from monthly data, index 2000=100. Source: FRED, http://research.stlouisfed.org/fred2

Nominal WageHourly Earnings (MEI), quarterly observations, index 2000=100. Source: OECD, www.oecd.org.

Oil ConsumptionOECD Countries and World Petroleum (Oil) Demand (consumption), yearly observa-tions, thousand barrels per day. Source:http://www.eia.doe.gov/.

Crude Oil ProductionU.S., Persian Gulf, OPEC and World Production, yearly observations, thousand barrels per day. Source:http://www.eia.doe.gov/.

Imports and Exports by CountryU.S. Imports and Exports by Country are available for the major trad-ing partners: Canada, China, France, Germany, Japan, Mexico and U.K. Monthly frequency: 1974:01-2007:12. Not seasonally adjusted. The data presented in the pictures are yearly averages. Source: FRED, http://research.stlouisfed.org/fred2/.

The real GDP has been computed dividing the GDP by the GDP deflator. The real wage has been computed dividing the hourly earnings by the CPI. The real oil price has been computed dividing the nominal oil price by the CPI.

(29)

C

Tables and Figures

Table 1

Percentage change in the Real Oil Price During Each Oil Shock Period

Q2:73 - Q1:74 Q4:78 - Q2:80 Q4:98 - Q4:00 Q4:01 - Q3:05

97% 79% 85% 103%

Table 2

Percentage change between 4 quarters after and 4 quarters before the oil shock

Q1:74 Q3:79 Q3:99 Q1:03 CPI Inflation 3.29 3.26 1.15 -0.35 GDP Defl. Inflation 3.44 0.46 0.63 0.50 Real GDP Growth -6.02 -3.29 0.41 1.75 Real Wage -2.26 -4.76 0.69 0.67 Wage Inflation 3.1753 -0.8838 0.4428 -0.9325 Table 3

Percentage change between 8 quarters after and 8 quarters before the oil shock

Q1:74 Q3:79 Q3:99 Q1:03 CPI Inflation 3.5680 2.9422 1.4251 0.5428 GDP Defl. Inflation 3.1880 1.5933 1.0106 0.5794 Real GDP Growth -5.0952 -3.1710 -1.3663 2.4373 Real Wage -1.7812 -5.4471 0.7128 1.6178 Wage Inflation 1.2753 0.7356 0.3578 -0.6904

(30)

MAGYAR NEMZETI BANK

Table 4

Oil Consumption (%) Growth Rates - 2001:2006. The data on which growth rates have been computed have been

downloaded from the Energy Information Administration and refer to Thousand Barrels per Day.

Region/Country (%) Growth Rate

North America 5.21

Central & South America 4.78

Europe 2.09

Eurasia 11.24

Middle East 24.75

Africa 16.55

Asia & Oceania 15.84

Total OECD 2.90

Total Non-OECD 19.79

Total World 9.33

China 46.43

Table 5

Oil Consumption by Country as % of World Consumption in 2006. Data on oil consumption have been downloaded

from the Energy Information Administration and refer to Thousand Barrels per Day. Each entry in the table reports the % of world consumption accounted for by each region/country.

Region/Country Country Consumption/World Consumption(%)

United States 24.45

North America 29.49

Central & South America 6.59

France 2.32 Germany 3.15 Italy 2.05 Spain 1.88 United Kingdom 2.16 Europe 19.39 Russia 3.32 Eurasia 4.97 Iran 1.99 Saudi Arabia 2.53 Middle East 7.29 Africa 3.58 China 8.51 India 3.04 Japan 6.10 South Korea 2.57

Asia & Oceania 28.69

Total OECD 58.30

(31)

APPENDIX C

Figure 1

(32)

MAGYAR NEMZETI BANK

Figure 2

(33)

APPENDIX C

Figure 3

(34)

MAGYAR NEMZETI BANK

Figure 4

Real Wage and Wage Inflation - Q1:1960 - Q1:2008

Figure 5

(35)

APPENDIX C

Figure 6

(36)

MAGYAR NEMZETI BANK

Figure 7

(37)

APPENDIX C

Figure 8

(38)

MAGYAR NEMZETI BANK

Figure 9

Impulse Responses: Oil Supply Shock

Figure 10

References

Related documents

Summary of mean (range) coded instances of communicative modalities and other coded categories according to narrative condition – Participant J ……… .... NS use of pointing

Centaurea aggregata Centaurea cadmea Centaurea wiedemanniana Centaurea aziziana Centaurea protogerberi Centaurea sterilis Centaurea deusta

Consolidation chemotherapy in acute myeloid leukemia (AML) aims at eradicating residual leukemic cells and mostly comprises high-dose cytarabine with or without the addition

• RQ2: Can metrics and feedback provided by UML-Ninja help the stakeholders (e.g., researchers, students, practitioners) to obtain a better assessment of the quality of UML models..

Table 11 also indicates that using a heterogeneous fleet of vehicles with a fixed speed of 100 km/h is better than using a homogeneous fleet of vehicles and optimizing their speeds

The objective of the optimization problem is to optimize the parameter of the time mul- tiplier setting (TMS) and plug setting (PS) in order to minimize the total operating time of

By additionally exploiting our information on the households’ knowledge about power prices, we also employ an Endogenous Switching Regression Model to estimate price elasticities

492-520; Micaela Frulli, ‘Advancing the Protection of Cultural Property through the Implementation of Individual Criminal Responsibility: the Case-Law of the International